In the mid-1980s, an article in a popular science magazine that talked about genetic engineering—recombinant DNA technologies, cloning, and related advances—caught my attention as a high schooler in São Paulo, Brazil. The story referenced a science book filled with these topics, written by scientists from the University of São Paulo.
This motivated me to approach the authors, microbiologists Elisabete Vicente and Beatriz Fernandes, in the interim between high school and my studies at the University of São Paulo to join their research group. Naturally, they didn’t let me run experiments at first. I started by washing glassware and gradually worked my way up to shadowing senior students. Then, when I was an undergraduate student there, I was able to conduct experiments.
My project involved transforming Saccharomyces cerevisiae to trigger a chromosomal recombination that would mutate and inactivate the radiation sensitive 52 (RAD52) gene, which plays a key role in DNA repair. This was long before kits and robots—everything was manual and full of painstaking steps. For months, I struggled to inactivate RAD52 in the yeast. I tried everything, from swapping out consumables one by one to changing the water source, but nothing worked.
I became increasingly frustrated with my failed experiments; it felt as if the universe was against me. At that time, I heard about the protein folding problem in one of my lectures. Because I was increasingly interested in computational biology, the topic felt like a natural fit. With the encouragement of my advisors, I left the biology lab and turned to physics. For my master’s degree, I joined physicist Vera Bohomoletz Henriques’s group at the same university that focused on simulations. Although experimental biology didn’t work out for me, it set me on a path toward computational biology and a deep fascination with the protein folding problem.
Even though AlphaFold has essentially solved much of this problem, it hasn’t eliminated the need to understand protein folding. Instead, current computational tools have opened the door to exploring the cellular biology of protein interactions at true atomic resolution. I’m still excited by that challenge.
Looking back, I see my departure from experimental work as a personal failure, but it ultimately steered me toward a direction I love. Today, I’m grateful for where that path led me—to my role as a computational biophysicist and research group leader at the University of Montreal.
This interview has been edited for length and clarity.
Do you have an Epic Fail story of your own? Share your best, worst mistake with us for a chance to be featured.